Abstract
“No great discovery was ever made without a bold guess.”
—Sir Isaac Newton (c. 1600s)
W
Cancer registry data have demonstrated a very high “familiality” in thyroid cancer diagnoses and thus, a strong genetic component. Relatedly, thyroid cancer is a component malignancy of multiple well-defined hereditary cancer syndromes, including FAP (175100; OMIM), Cowden syndrome (158350; OMIM), Carney complex type I (160980; OMIM), DICER1 syndrome (606241; OMIM), and Werner syndrome (277700; OMIM). Compared with a 1% lifetime risk of thyroid cancer in the general population, individuals at genetic risk carry markedly increased lifetime risks, often early onset, multifocal, and aggressive, and importantly, almost always carry risks of extra-thyroidal cancers. Importantly, whether hereditary tumors behave similarly to their sporadic counterparts at the molecular and cellular levels also remains largely uncharacterized—often limited by the rarity of the syndromic forms of the disease.
It is in these contexts that the article by Nieminen et al. is very timely (1). The authors studied FAP, a hereditary cancer predisposition disorder caused by germline adenomatous polyposis coli (APC; 611731; OMIM) loss-of-function mutations. In addition to colonic polyposis and heightened colorectal cancer risks, individuals with FAP have up to 12% lifetime risk of PTC compared with the baseline risk of 1% in the general population (2). However, the molecular determinants of the association between FAP and PTC are poorly understood, with contradictory data regarding the role of “second-hit” somatic APC mutations in the process of thyroid carcinogenesis. Hence, the authors performed whole-genome sequencing on a set of 12 paired thyroid tumor and normal DNA from FAP-PTC patients. All patients harbored germline APC mutations, consistent with an FAP diagnosis.
Interestingly, the somatic mutation landscape revealed that 7/12 (58%) of FAP-PTC tumors carried second-hit somatic APC mutations, with 1 tumor having clear loss of heterozygosity (LOH) of APC and 2 other tumors demonstrating putative APC LOH. Furthermore, they identified somatic mutations in KMT2D (lysine-specific methyltransferase 2D; 602113; OMIM) in 7/12 (58%) of FAP-PTC tumors. While the authors identified multiple other recurrent somatically mutated genes in the FAP-PTC tumors, the landscape of somatic mutations was markedly distinct from that of sporadic PTC. First, somatic APC and KMT2D mutations are exceedingly rare (<3%) in sporadic PTC (3). Nonetheless, there is at least one report of a biallelic somatic APC mutation in a sporadic (non-FAP) cribriform-morular variant of an early-onset PTC (4), bolstering the evidence that biallelic dysfunction of APC is important in the pathogenesis of the cribriform-morular variant of PTC.
Conversely, somatic BRAF (164757; OMIM) mutations, the vast majority of which consist of the driver BRAF p.V600E mutation, occur in about 60% of all sporadic PTC (3). However, the authors identified the BRAF p.V600E mutation in only 2/12 (17%) of the FAP-PTC tumors, a remarkable difference. Notably, BRAF and APC somatic mutations were mutually exclusive in these tumors. Further supporting an altered somatic mutation landscape is the absence of somatic mutations in NRAS, HRAS, and KRAS, as well as RET (164761; OMIM) translocations/inversions (including RET/PTC), which are also well-known driver alterations in sporadic PTC (3).
The FAP-PTC somatic landscape in and of itself is intriguing. In cancer evolution theory, it is widely accepted that only a distinct portion of somatic mutations function as drivers that confer cell fitness and are positively selected over the course of tumor formation (5). In this study, while the somatic APC and BRAF driver mutations were mutually exclusive, KMT2D, KMT2C (lysine-specific methyltransferase 2c; 606833; OMIM), and other reported somatic variants tended to co-occur with the somatic driver APC and BRAF mutations. These data suggest that at least KMT2D and KMT2C could have a context-dependent dual role as driver or passenger genes when mutated.
In the tumors harboring the BRAF p.V600E mutation, PTCs arising in the background of germline APC mutations could be said to share the same driver mutations as sporadic PTCs. However, it is also tempting to speculate whether those PTCs with the BRAF p.V600E mutation occurred sporadically (i.e., incidentally and unrelated to FAP), particularly because the thyroid cancer diagnoses occurred at relatively older ages (44 and 62 years). Indeed, similar to the other demographic and clinical correlates that the authors discussed, this will be difficult to ascertain with the small sample size of patients, inherent to the majority of such rare hereditary cancer predisposition syndromes.
In the sporadic thyroid cancer arena, multiregion sequencing of individual tumors recently revealed that intratumor heterogeneity is relatively common in PTC (6). Importantly, while the majority (68%) of tumors followed a linear evolution pattern, the remaining (32%) tumors followed a branched evolution pattern, leading to divergent subclones sharing only partial sets of mutations. Notably, it was estimated that 89% of the branched subclones would have appeared falsely clonal in the absence of multisampling of individual tumors. Moreover, PTC patients with a higher burden of subclonal mutations had a significantly higher risk of relapse compared with those with a lower burden. In the case of FAP, it remains tantalizing to characterize the basis for such a remarkably different somatic mutation landscape. Indeed, identifying such detailed molecular correlates are predicted to have important biological implications toward understanding, at a higher resolution, determinants of PTC initiation and progression with an APC-FAP host background.
After characterization of the somatic tumor mutation landscape, the authors then examined the occurrence of germline variants in genes known to be associated with familial nonmedullary thyroid cancer (FNMTC) (7). They identified 9/12 (75%) FAP-PTC patients carrying germline variants in 17 different genes (1). Interestingly, multiple RNF213 (613768; OMIM) variants were identified in three FAP-PTC patients. It is important to note that the authors focused on rare (minor allele frequency <0.01 in gnomAD) and nonsynonymous variants. Therefore, in the absence of characterization of the full germline mutation spectrum, these data provide but the tip of the iceberg for the possible contribution of FNMTC-associated germline variants to the etiology of thyroid cancer in FAP.
Although beyond the scope of the presented study, two aspects warrant further examination in the germline context. First, whether the studied FAP-PTC patients had a family history of PTC and/or other FAP-associated cancers. Second, what naturally follows is whether the germline APC mutations were inherited or de novo. Approximately 20–25% of individuals with FAP harbor de novo APC mutations (2). Together with the genetic data, the clinical phenotypic data will be powerful to dissect germline predisposing versus modifying factors for PTC occurrence in particular individuals within particular families.
Considered the fastest rising incident cancer in women for the past decade and a half, PTC is the most prevalent endocrine malignancy (8). The finding that FAP-associated PTCs are characterized by a distinct somatic mutation landscape compared with sporadic PTCs suggests equally distinct processes to account for tumor initiation in this context. At the end of the day, these authors may have given us the first clues to the link between biallelic-APC or somatic KMT2D/KMT2C dysfunction and the cribriform-morular variant of PTC, whereas, BRAF-associated FAP-PTC may be more akin to classic PTC.
“The possession of knowledge does not kill the sense of wonder and mystery. There is always more mystery.”
—Anaïs Nin (c. 1900s)
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received.